My model (partial code):
class Observation(models.Model):
date = models.DateField()
geom = models.PointField()
values = models.ManyToManyField(Label, through='Value', null=True, blank=True)
objects = models.GeoManager()
class Value(models.Model):
observation = models.ForeignKey(Observation)
label = models.ForeignKey(Label)
value = models.CharField(max_length=100)
objects = models.GeoManager()
When I manage an Observation object in the admin interface, it still says at least one value per observation is required.
Am I doing something wrong, is this a bug, or should I write a derived Admin class to solve this?
I solved this by improving my ERM. The field values in Observation is obsolete, since you get a value_set from the ForeignKey relation in Value.
Still a weird side effect, but since there were no replies I'll consider it something that doesn't occur often.
This is happened to me too.
How exactly you got it resolved?
Following modification seems to do the trick:
But not sure, what effect it's having at DB level.
class Value(models.Model):
label = models.ForeignKey(Label, blank=True, null=True)
Related
I have two models
class Customer(models.Model):
name = models.CharField(max_length=255, unique=True)
default_contact = models.ForeignKey("CustomerContact", verbose_name="...", related_name="default_contacts", null=True, on_delete=models.SET_NULL)
etc.
And
class CustomerContact(models.Model):
customer = models.ForeignKey(Customer, related_name='contacts')
user = models.OneToOneField(User, related_name='user_contacts', on_delete=models.SET_NULL)
address = models.ForeignKey(CustomerAddress, ....)
In this example Customer points to CustomerContact.
At the same time CustomerContact points to Customer.
My coworker says that pointing Customer points to CustomerContact violates the OneToMany nature of ForeignKey.
What am I doing wrong?
As I can see you want to have many CustomerContact related to one Customer, but Customer can also pick his favourite (or one can be set by manager). It's valid approach.
It can go both ways, as long as you will secure related_names properly.
default_contact = models.ForeignKey("CustomerContact", ... related_name="default_contacts") <= should be changed
default_contacts should be changed (i.e. to default_for_customers) because this name is for reversed relation, so actually for CustomerContact. It means, that you can use it in the following situation:
cc = CustomerContact.objects.get(id=1)
cc.default_for_customers.all() <= this will return QuerySet of Customer objects that is default for
It's simplier and less confusing.
Having a problem with a M2M related queryset:
simplified model is
class Category(models.Model):
name = models.CharField(max_length=200)
class Product(models.Model):
name = models.CharField(max_length=200)
category = models.ManyToManyField(
Category, blank=True, null=True, related_name="categorytest")
I usually do related queries using related name (which works fine here no problem) yet I wanted to try "set_all()" in this case.
Say I do:
elecs = Category.objects.get(name="Electronics")
then:
elecs.product_set.all()
I get AttributeError: 'Category' object has no attribute 'product_set'. I am a bit surprised because this is an error that happens when using a queryset instead of a model instance yet, if a confirmation was needed, a type(elecs) gives me <class 'store.models.Category'>.
I migrated/erased db multiple times, no change.
Any idea ? Thanks
You specify the name of the relation in reverse with the related_name=… parameter [Django-doc]. Since you set related_name='categorytest', you thus access these with:
elecs.categorytest.all()
But likely using categorytest is not a good idea. It might be better to specify this for example as related_name='products'.
I'm trying to access the grandchildren records in a list to avoid duplicate records. In this example, a tag can only be used once across articles for a given author. I will use the resulting list of grandchildren records in my clean function to return validation errors.
class Article(models.Model):
tag = models.ManyToManyField(Tag)
author = models.ForeignKey(Author, on_delete=models.CASCADE)
class Tag(models.Model):
class Author(models.Model):
Right now I can do this:
print(author.articles.first().tag.first())
Travel
I'd like to be able to use something like author.articles.tags.all() to return the list and check the submitted form against it to raise a ValidationError message to the user.
How can this be done efficiently with the basic Many-to-Many setup without creating an intermediate table for the tag relationships? This is solely in the Admin interface, in case that matters at all.
i come from the link you posted on upwork,
the way i understand your question,
what you want to achieve seems to be impossible ,
what i think can work , is to fetch articles related to the author, with their corresponding tags,
after that they are retrieved you do filtering and remove duplicates.
otherwise the tag has nothing to connect it with the author,,
I'm so jealous that's a great answer #since9teen94
Be aware, I will not base my answer in the easiest solution but how we model reality (or how we should do it).
We put tags after things exists right?
In that order I will make your life horrible with something like this:
from django.db import models
class Author(models.Model):
name = models.CharField(max_length=30, null=False)
class Article(models.Model):
name = models.CharField(max_length=30, null=False)
author = models.ForeignKey(Author, on_delete=models.CASCADE)
class Tag(models.Model):
name = models.CharField(max_length=30, null=False, unique=True)
articles = models.ManyToManyField(Article)
but believe me you don't want to follow this approach it will make your work harder.
You can search for Articles based on Tags directly
Tag.objects.filter(name='Scify').first().articles.all() # name is unique
The real issue with this is that the reverse lookup is really complex (I mean.. get ready for debug)
Article.objects.filter(
id__in=list(
Tag.objects.filter(name='Scify').first().articles.values_list('id', flat=True)
)
)
I am sure this does not solve your problem and I don't like complex code for no reason but if you're open to suggestions I don't mind to think different and add some options
Edit:
About the author and clean repeated tags.. well you don't have to deal with that and if you want to find all Tag your author has you could loop the tags
for tag in Tag.objects.all():
if tag.articles.filter(author__name='StackoverflowContributor'):
print(tag.name)
# > Scify
I am just saying that there are options not that this is the best for you but don't be afraid of the terminal, it's really cool
The Django ORM is pretty cool when you get used to it, but regular SQL is pretty cool too
# created_at, updated_at, name and tag_val are variables I
# added due to my slight ocd lol
class Author(models.Model):
name = models.CharField(max_length=255)
created_at = models.DateTimeField(auto_now_add=True)
updated_at = models.DateTimeField(auto_now=True)
class Tag(models.Model):
tag_val = models.CharField(max_length=255)
created_at = models.DateTimeField(auto_now_add=True)
updated_at = models.DateTimeField(auto_now=True)
class Article(models.Model):
author = models.ForeignKey(Author,related_name='articles', on_delete=models.CASCADE)
tags = models.ManyToManyField(Tag, related_name='articles')
created_at = models.DateTimeField(auto_now_add=True)
updated_at = models.DateTimeField(auto_now=True)
I can write my query like this, assuming the variable 'author' has been assigned an Author object instance, and get a list of dictionaries [{'tags':1},{'tags':2}] where the value is the auto generated primary key id of the Tag object instance
author.articles.values('tags').distinct()
I have a Notes and a NoteRefs fields where the NoteRefs has a foreign key to the Notes. I need to query the Notes but order by the related field (ie. the NoteRefs' start_ref field).
How might I do that through the django ORM? Here's kinda what works in SQL
SELECT
note.user_id,
note.content,
note.created,
note.modified
FROM noteref
INNER JOIN note
ON note.id = noteref.note_id
ORDER BY noteref.start_ref
I can't use Note.order_by('related_field'), because the related field isn't part of the Note Model. From what I can tell, that seems to be what the documentation says to do. How can I sort on the related field here?
EDIT: Model information
class Note(models.Model):
user = models.ForeignKey(User, db_index=True)
content = models.TextField()
created = models.DateTimeField(auto_now_add=True)
modified = models.DateTimeField(auto_now=True)
class NoteRef(models.Model):
note = models.ForeignKey(Note, db_index=True)
_order = models.IntegerField(default=0)
book = models.IntegerField(max_length=2)
start_ref = models.IntegerField(max_length=8, db_index=True)
end_ref = models.IntegerField(max_length=8, db_index=True)
ref_range = models.IntegerField()
passage = models.CharField(max_length=50)
You should try Note.objects.order_by("noterefs__start_ref")
The documentation doesn't make this very clear, as it uses a ForeignKey to self, but it works.
Now, the docs also warn against the possibility of duplicate objects showing up if you have multiple NoteRefs for a single Note, so you should double-check this.
I have just begun to play around with Django admin views, and to start off, I am trying to do something very simple: showing several fields in the listing of objects using list_display as explained here: https://docs.djangoproject.com/en/dev/ref/contrib/admin/
This is my dead simple code:
class ArticleAdmin(admin.ModelAdmin):
list_display = ('title', 'category')
Unfortunately, the list_display option is causing the columnar view to appear, but only some of the objects (40 out of 85) are now displaying in the listing. I cannot deduce why certain objects are showing over the others - their fields look like they are filled similarly. It's clearly not paginating, because when I tried it on an admin of another model, it showed only 2 objects out of about 70 objects.
What might be going on here?
[UPDATE] Article Model:
class Article(models.Model):
revision = models.ForeignKey('ArticleRevision', related_name="current_revision")
category = models.ForeignKey('meta.Category')
language = models.ForeignKey('meta.Language', default=get_default_language)
created = models.DateTimeField(auto_now_add=True, editable=False)
changed = models.DateTimeField(auto_now=True, editable=False)
title = models.CharField(max_length=256)
resources = models.ManyToManyField('oer.Resource', blank=True)
image = models.ManyToManyField('media.Image', blank=True)
views = models.IntegerField(editable=False, default=0)
license = models.ForeignKey('license.License', default=get_default_license)
slug = models.SlugField(max_length=256)
difficulty = models.PositiveIntegerField(editable=True, default=0)
published = models.NullBooleanField()
citation = models.CharField(max_length=1024, blank=True, null=True)
Before adding list_display:
After adding list_display:
[UPDATE] This behaviour occurs only when ForeignKey fields are included in list_display tuple. Any of them.
[UPDATE] Category model code:
class Category(models.Model):
title = models.CharField(max_length=256)
parent = models.ForeignKey('self')
project = models.NullBooleanField(default=False)
created = models.DateTimeField(auto_now_add=True, editable=False)
slug = models.SlugField(max_length=256, blank=True)
def __unicode__(self):
return self.title
This behavior is caused by a foreign key relation somewhere that is not declared as nullable, but nonetheless has a null value in the database. When you have a ManyToOne relationship in list_display, the change list class will always execute the query using select_related. (See the get_query_set method in django.contrib.admin.views.ChangeList).
select_related by default follows all foreign keys on each object, so any broken foreign key found by this query will cause data to drop out when the query is evaluated. This is not specific to the admin; you can interactively test it by comparing the results of Article.objects.all() to Article.objects.all().select_related().
There's no simple way to control which foreign keys the admin will look up - select_related takes some parameters, but the admin doesn't expose a way to pass them through. In theory you could write your own ChangeList class and override get_query_set, but I don't recommend that.
The real fix is to make sure your foreign key model fields accurately reflect the state of your database in their null settings. Personally, I'd probably do this by commenting out all FKs on Article other than Category, seeing if that helps, then turning them back on one by one until things start breaking. The problem doesn't have to be with a FK on an article itself; if a revision, language or category has a broken FK that will still cause the join to miss rows. Or if something they relate to has a broken FK, etc etc.